feat(transformers/models): add models of d_fine, granitemoehybrid and efficientloftr

mindone/transformers/__init__.py

@@ -373,6 +373,7 @@
 )
 from .models.ctrl import CTRLForSequenceClassification, CTRLLMHeadModel, CTRLModel, CTRLPreTrainedModel
 from .models.cvt import CvtForImageClassification, CvtModel, CvtPreTrainedModel
+from .models.d_fine import DFineForObjectDetection, DFineModel, DFinePreTrainedModel
 from .models.dac import DacModel, DacPreTrainedModel
 from .models.data2vec import (
     Data2VecAudioForAudioFrameClassification,
@@ -463,6 +464,7 @@
     DPRReader,
 )
 from .models.dpt import DPTForDepthEstimation, DPTImageProcessor, DPTModel, DPTPreTrainedModel
+from .models.efficientloftr import EfficientLoFTRForKeypointMatching, EfficientLoFTRModel, EfficientLoFTRPreTrainedModel
 from .models.efficientnet import (
     EfficientNetForImageClassification,
     EfficientNetImageProcessor,
@@ -650,6 +652,7 @@
 )
 from .models.granite import GraniteForCausalLM, GraniteModel, GranitePreTrainedModel
 from .models.granitemoe import GraniteMoeForCausalLM, GraniteMoeModel, GraniteMoePreTrainedModel
+from .models.granitemoehybrid import GraniteMoeHybridForCausalLM, GraniteMoeHybridModel, GraniteMoeHybridPreTrainedModel
 from .models.granitemoeshared import GraniteMoeSharedForCausalLM, GraniteMoeSharedModel, GraniteMoeSharedPreTrainedModel
 from .models.grounding_dino import (
     GroundingDinoForObjectDetection,
@@ -666,6 +669,7 @@
     HeliumModel,
     HeliumPreTrainedModel,
 )
+from .models.hgnet_v2 import HGNetV2Backbone, HGNetV2ForImageClassification, HGNetV2PreTrainedModel
 from .models.hiera import (
     HieraBackbone,
     HieraForImageClassification,
@@ -1190,7 +1194,7 @@
     RoFormerModel,
     RoFormerPreTrainedModel,
 )
-from .models.rt_detr import RTDetrForObjectDetection, RTDetrModel, RTDetrPreTrainedModel
+from .models.rt_detr import RTDetrForObjectDetection, RTDetrImageProcessor, RTDetrModel, RTDetrPreTrainedModel
 from .models.rt_detr_v2 import RTDetrV2ForObjectDetection, RTDetrV2Model, RTDetrV2PreTrainedModel
 from .models.rwkv import RwkvForCausalLM, RwkvModel, RwkvPreTrainedModel
 from .models.sam import SamImageProcessor, SamModel, SamPreTrainedModel, SamProcessor

mindone/transformers/models/__init__.py

@@ -54,6 +54,7 @@
     convnextv2,
     ctrl,
     cvt,
+    d_fine,
     dac,
     data2vec,
     dbrx,
@@ -67,6 +68,7 @@
     distilbert,
     dpr,
     dpt,
+    efficientloftr,
     efficientnet,
     electra,
     emu3,
@@ -99,8 +101,10 @@
     gptj,
     granite,
     granitemoe,
+    granitemoehybrid,
     granitemoeshared,
     groupvit,
+    hgnet_v2,
     hiera,
     hubert,
     idefics,

mindone/transformers/models/auto/configuration_auto.py

@@ -124,6 +124,7 @@
         ("granitemoeshared", "GraniteMoeSharedConfig"),
         ("groupvit", "GroupViTConfig"),
         ("helium", "HeliumConfig"),
+        ("hgnet_v2", "HGNetV2Config"),
         ("hiera", "HieraConfig"),
         ("hubert", "HubertConfig"),
         ("ibert", "IBertConfig"),
@@ -392,6 +393,7 @@
         ("granitemoeshared", "GraniteMoeSharedMoe"),
         ("groupvit", "GroupViT"),
         ("helium", "Helium"),
+        ("hgnet_v2", "HGNet-V2"),
         ("hiera", "Hiera"),
         ("hubert", "Hubert"),
         ("ibert", "I-BERT"),

mindone/transformers/models/auto/image_processing_auto.py

@@ -73,6 +73,7 @@
             ("owlv2", ("Owlv2ImageProcessor",)),
             ("owlvit", ("OwlViTImageProcessor",)),
             ("qwen2_5_vl", ("Qwen2VLImageProcessor",)),
+            ("rt_detr", ("RTDetrImageProcessor",)),
             ("sam", ("SamImageProcessor",)),
             ("segformer", ("SegformerImageProcessor",)),
             ("siglip", ("SiglipImageProcessor", "SiglipImageProcessorFast")),

mindone/transformers/models/auto/modeling_auto.py

@@ -118,6 +118,7 @@
         ("groupvit", "GroupViTModel"),
         ("grounding-dino", "GroundingDinoModel"),
         ("helium", "HeliumModel"),
+        ("hgnet_v2", "HGNetV2Backbone"),
         ("hiera", "HieraModel"),
         ("hubert", "HubertModel"),
         ("ibert", "IBertModel"),
@@ -583,6 +584,7 @@
         ("dinov2_with_registers", "Dinov2WithRegistersForImageClassification"),
         ("efficientnet", "EfficientNetForImageClassification"),
         ("focalnet", "FocalNetForImageClassification"),
+        ("hgnet_v2", "HGNetV2ForImageClassification"),
         ("hiera", "HieraForImageClassification"),
         ("ijepa", "IJepaForImageClassification"),
         ("imagegpt", "ImageGPTForImageClassification"),
@@ -1212,6 +1214,7 @@
         ("dinov2", "Dinov2Backbone"),
         ("dinov2_with_registers", "Dinov2WithRegistersBackbone"),
         ("focalnet", "FocalNetBackbone"),
+        ("hgnet_v2", "HGNetV2Backbone"),
         ("hiera", "HieraBackbone"),
         ("maskformer-swin", "MaskFormerSwinBackbone"),
         ("pvt_v2", "PvtV2Backbone"),

mindone/transformers/models/d_fine/__init__.py

@@ -0,0 +1,14 @@
+# Copyright 2025 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+from .modeling_d_fine import *